The only assumption I know of is that the half-life of any given radioisotope is assumed to be constannt.
As far as I know there have never been two samples of the same radioisotope found to have different half-lives, so the assumption seems sound.
Do you have evidence that the half lives can be made variable to account for a margin of error of 4500000% ?
Yet you have no explanation for observed errors nor even what percent of these perceived errors are discarded.
Worst of all for critical thinking and the scientific method ~ zero explanation for the far easier to observe natural clocks that I’ve been pointing out throughout this thread.
Another couple of assumptions for radioisotope dating:
Initial ratio of father and daughter elements, and
Volcanic heat resets these radioisotope clocks.